It has been indicated that unnatural hetero-aggregates form spontaneously between host cell-derived contaminants and recombinant proteins produced by in vitro cell culture methods (Shukla et al., Biotechnol. Progr. (2008) 24:1115-1121; Luhrs et al., J. Chromatogr. B (2009) 877:1543-1552; Mechetner et al., J. Chromatogr. B (2011) 879:2583-2594; Gagnon et al., J. Chromatogr. A, (2011) 1218:2405-2412; Gagnon, Bioprocessing J. (2010) 9(4):14-24). These hetero-aggregates may be considered unnatural in two respects: 1) constituent contaminants are often of non-human origin, secreted by living non-human host cells or released into the culture media when non-human host cells lyse upon death. In living humans, such non-human contaminants do not exist; and 2) constituent contaminants accumulate to high concentrations in comparison to human in vivo systems where dead cell constituents are quickly eliminated. Accordingly, recombinant products are exposed to high levels of strongly interactive contaminants at concentrations that typically do not occur in living systems. Meanwhile, high expression levels of recombinant proteins make them suitable substrates for non-specific associations with these non-human contaminants, favoring the formation of undesirable hetero-aggregates of diverse composition.
The contaminating protein content of hetero-aggregates has been addressed to some extent via direct targeting of the contaminating protein (Shukla et al. and Gagnon et al. supra), as well as indirectly via targeting of the corresponding DNA component responsible for the contaminating protein (Luhrs et al. and Gagnon supra). A reduction of antibody aggregate level has been indicated when some complexes are dissociated (Shukla et al., Mechetner et al., and Gagnon supra). The ability of anion exchangers to reduce levels of antibody-contaminant complexes has been disclosed (Luhrs et al. and Gagnon et al. supra), but an anion exchange treatment that was able to fully eliminate hetero-aggregates has not been indicated. Size exclusion, cation exchange, and hydrophobic interaction chromatography have also been employed in attempts to reduce hetero-aggregates, but these techniques were generally inferior to anion exchange (Gagnon et al. supra).
The specific source of contaminants that form stable associations with antibodies is not always known (see, for example, Shukla et al. supra). Some efforts have focused on DNA contaminants with little attention to the specific source of other possible contaminants (Gagnon et al. and Gagnon supra). Some efforts indicating an association of host contaminants with aggregates in antibody preparation have focused specifically on contaminants comprising chromatin catabolites (Luhrs et al. and Mechetner et al. supra). In these examples, aggregation may be mediated directly through the immunospecificity of the antibody for chromatin catabolites such as histones and DNA. It has been indicated that chromatin catabolites are also capable of forming stable complexes with antibodies via non-specific interactions. Thus, monoclonal antibodies with known immunospecificities for antigens not including chromatin catabolites, can form highly stable aggregates of diverse descriptions with nucleosomes, histones, and DNA derived from the nuclei of dead host cells. It has been particularly indicated that chromatin catabolites are highly represented in high molecular weight (HMW) aggregates. HMW aggregates are of particular concern because of their suspected involvement in promoting the formation of therapy-neutralizing antibodies. HMW aggregates are generally defined as aggregates of a size greater than small multiples of the antibody of interest. For example, 2-antibody associations are not considered HMW aggregates, nor are most 4-antibody aggregates. However, aggregates of much greater size, such as corresponding to about 8 to about 10 or more antibodies may be generally classified as HMW aggregates.
Treating antibody preparations with agents that might be expected to dissociate hetero-aggregates has generally proven ineffective. For example, employing high concentrations of urea, salts, or combinations of the two does not substantially dissociate IgM-contaminant hetero-aggregates (Gagnon et al. supra). Protein A affinity chromatography with pre-elution washes of urea, alcohol, and surfactants has been indicated to reduce hetero-aggregate levels more effectively than without washes (Shukla et al. supra), as did pre-elution washes combining urea, salt, and EDTA with protein G affinity chromatography (Mechetner et al. supra). Anion exchange chromatography with a pre-elution wash of urea has been indicated to reduce hetero-aggregates more effectively than in the absence of a urea wash (Gagnon et al. supra). Cation exchange chromatography has also been indicated to reduce hetero-aggregates more effectively with a pre-elution EDTA wash than without the wash (Gagnon et al. supra). Finally, hydroxyapatite with pre-elution washes of urea and/or salt have also reduced hetero-aggregates more effectively than without such washes (Gagnon supra). Despite these observations, in general, the use of dissociating agents in pre-elution washes of antibodies bound to chromatography columns has been only moderately successful.
Organic multivalent cations have been indicated for the precipitation of acidic proteins (Farhner et al., U.S. Patent Application No. 20080193981; Ma et al., J. Chromatogr. B (2010) 878:798-806; Peram et al., Biotechnol. Progr., (2010) 26:1322-1326; Glynn, in U. Gottschalk (ed.), Process Scale Purification of Antibodies, J. T. Wiley and Sons, (2009) Hoboken, 309-324), as well as for precipitation of DNA and endotoxins (Glynn supra; Cordes et al., Biotechnol. Progr., (1990) 6:283-285; Dissing et al., Bioseparation, (1999) 7 221:9-11) and inactivation of virus (Bernhardt, U.S. Pat. No. 5,559,250). Multivalent metal cations have also been indicated to remove DNA and endotoxin from some protein preparations (Akcasu et al., Nature, (1960) 187:323-324; Matsuzawa et al., Nucl. Acids Res., (2003) 3(3):163-164; Christensen et al., Prot. Expr. Purif., (2004) 37:468-471; Kejnovsky et al., Nucl. Acids Res., (1997) 25:1870-1871; Ongkudon et al., Anal. Chem., (2011) 83 391:13-17).